Glioblastomas are aggressive and invasive brain tumors that generally lead to death within a year of diagnosis. No cure is available, and current treatments prolong life by only a few months, often at the expense of quality of life. A number of parvoviruses have been reported to show oncolytic potential against cancer cells, and a rat parvovirus, H-1, is currently in clinical trials to treat glioma patients. In our initial work screning a large number of parvoviruses, we found one relatively obscure parvovirus, LuIII, that performed substantially better than any of the others tested, including H-1, and was the only parvovirus tested that successfully killed multiple human gliomas. LuIII appears safe in the brain, and does not target or kill neurons and shows minimal infection of normal glia. Here we test a number of hypotheses related to the ability of LuIII to selectively infect, replicate in, an kill human glioma. In the first set of experiments, we test the hypothesis that LuIII will successfully target and kill glioblastoma cells that are transplanted into the brain, with relativey little toxicity to the normal brain, both after an intratumoral virus injection, and after intravenus inoculation. Tumors are detected by expression of a red fluorescent reporter, and virally infected cells are detected by green immunofluorescent staining for the LuIII viral antigen NS1. A critical clinical problem with glioma is their recurrence after surgical or radiation treatment; here we tes the hypothesis that LuIII may maintain an asymptomatic low level of infection in normal human glial cells, and that LuIII arising from normal cells will thereby attenuate or block glioma cell expansion or recurrence. We will corroborate our preliminary findings and test the hypothesis that LuIII does not generate an anti-viral interferon (IFN) response, and that LuIII infection is nt attenuated by IFN. This independence from the IFN system sets LuIII apart from a number of other oncolytic viruses used in the brain which are sensitive to IFN. LuIII's insensitivity to IFN would allow a co-treatment with LuIII and IFN (IFN is a partially effective anti-tumor treatment in a subset of brain tumors), or co-treatment with LuIII together with an IFN-sensitive oncolytic virus. We will employ deep whole exome genetic sequencing, and sequencing of the mRNA transcriptome, to search for gene mutations in glioma that correlate with high levels of LuIII infection. This genetic information is useful both to understand the mechanisms underlying LuIII's selective infection of gliomas, and also may prove useful as a diagnostic predictor of which tumor-related mutations are most likely to be associated with a high LuIII infection. Finally, our preliminary data suggest that LuIII is effective at not only targeting and killing glioblastoma, but also infects and kills other cancers that invade the brain, including melanoma and lung cancer. Lung cancer metastasis is the most common problem involving secondary cancer in the brain. We will test the hypothesis that LuIII targets lung cancer cells in the brain n experiments parallel to those above involving glioma. Our central goal is to test the potential of LuIII as a safe and effective means of substantially attenuating or destroying brain tumors in humans.